The biggest roadblock to AI adoption is a lack of skilled workers

by JIM HARE, GARTNER, INC., Nov 4, 2017

While there is nearly universal agreement that artificial intelligence offers the promise of revolutionary benefits, recent survey findings from Gartner reveal almost 60 percent of organizations surveyed have yet to take advantage of the benefits of AI. Perhaps even more surprisingly, only a little more than 10 percent of surveyed businesses have deployed or implemented any AI solution at all.

Based on the survey, there appears to be a gap between AI’s promise and the ability for an enterprise to implement it. A further confirmation of that point is the finding that close to half of the surveyed organizations state they prefer to buy pre-packaged AI solutions or use AI capabilities already embedded in their applications. This shouldn’t be a surprise as end-user organizations are looking to use AI to help better solve business problems. They aren’t looking to simply buy AI technologies as an end in itself.

A vital factor is driving the preference for pre-packaged AI or AI-embedded applications. Many businesses aren’t prepared to enact a custom solution themselves due to a lack of in-house skills.

Gartner’s analysis has concluded that the skills gap is the most significant barrier to AI adoption.

Respondent organizations are in the very early stages of their AI projects. In fact, Gartner discovered that many organizations are still struggling to move from descriptive analytics toward foundational machine learning solutions for predictive and prescriptive analytics.

Another discovery from our research is that the organizations implementing AI solutions are not just those that label themselves as “aggressive,” meaning they welcome using cutting-edge technologies. In fact, more than half of our survey participants that reported implementing AI solutions label their organizations as “mainstream” — organizations that typically wait for technologies to mature.

AI: Still in the knowledge-gathering stage

Businesses have a strong interest in AI — inquiries from our clients to discuss AI topics have quadrupled from 2015 to 2017. In January 2016, the term, “artificial intelligence” didn’t even make it into our top 100 search terms. A year later, it ranked at No. 11, and in May 2017 it was at No. 7. This provides proof there is keen interest in understanding how AI can be used as part of a digital business strategy.

That said, about one-third of the survey respondents claim to face challenges in defining their AI strategy. This makes sense, given that 59 percent of organizations are still in the knowledge-gathering stage. Security and integration were also reported as challenges (30 percent and 27 percent, respectively), primarily by organizations in the knowledge gathering-strategy development stage.

Surprisingly, determining how to measure value from using AI was only a challenge for 23 percent of our respondents. This is likely because the majority of these organizations are still developing their strategies and don’t yet understand the importance of measuring the business value of the solution under consideration.

Time to go to school

We found that while organizations continue to have difficulties finding experienced data scientists for advanced analytics projects, it’s even harder to find employees skilled in AI techniques such as deep learning.

Much of AI innovation is happening at the university level, and the graduating students are joining cloud AI providers such as Google, Amazon, and Microsoft or launching their own startups to take advantage of investments from the venture capital community.

Many businesses are therefore seeking to update their in-house skills. Some organizations are also engaging system integrators with the goal of transferring knowledge from these system integrators to their own data scientists.

Enterprises should hire skilled students from local universities with AI specialization or project/internship experience. Ideally, they should look for students about to receive their BS and MS degrees in advanced data science and machine learning. They also should emphasize staff retraining and use rapid prototyping as a way to not only build team skills but also to also showcase the benefits of AI to upper management.

Build your strategy

Organizations should start building an AI strategy by partnering with in-house business executives to identify use cases where AI could be put to use — primarily to improve decision making and make processes more efficient. An organization should apply metrics to its AI initiative before beginning the pilot phase.

Once in production, the organization should continue using metrics as a way to refine and optimize the AI solution. They should also proactively communicate the metrics to senior management to prove ROI. This will be critical to obtaining buy-in from management. Organizations should also be sure to evaluate existing applications to understand plans to incorporate AI capabilities into other organizational solutions.

AI allows organizations to add intelligence to applications, services, and digital resources. Application leaders must establish when to use AI and how to address the challenges it will present to customers and employees.